Endoscopic Ultrasound-Guided Fine Needle Aspiration Cytology of Pancreatic Adenocarcinomas Revisited. A Detailed Cytological Analysis.

BACKGROUND: Early detection of pancreatic adenocarcinomas is essential for improving survival. In this regard, endoscopic ultrasound-guided fine-needle aspiration cytology (EUS-FNAC) has established itself as the method of choice for its ability to target lesions smaller than those which could be targeted by the traditional imaging methods like transabdominal ultrasound. Identifying these tumors correctly on FNA may be challenging because pancreatic adenocarcinomas may show a wide range of morphological features and the presence of contaminants from the gastrointestinal tract may show up as potential pitfalls. This study presents detailed cytomorphological analyses of 59 cases reported as pancreatic adenocarcinomas on smears and cell blocks. The clinical and histopathology follow-up data wherever available have also been presented.
MATERIALS AND METHODS: EUS-FNAC smears and cell blocks from cases reported as pancreatic adenocarcinomas were retrospectively evaluated with individual assessments of a range of features related to cellularity, cellular arrangement, cytoplasmic qualities, and nuclear features. Aspirates from peripancreatic lymph nodes, histopathology sections, and clinical records were reviewed wherever available.
RESULTS: Nonneoplastic cells like pancreatic ductal cells and acinar cells, duodenal, and gastric epithelia were detected along with neoplastic cells showing a wide range of variations in different cytomorphological characters. Often, a mixture of features was noted in the same case. Cell block preparations served as useful adjuncts since they made it possible to render unequivocal diagnoses of malignancies in cases where smears were hypocellular.
CONCLUSION: The study creates a useful knowledge base of cytomorphological features of pancreatic adenocarcinomas. Copyright:

INTRODUCTION

Among pancreatic tumors, adenocarcinomas are the most common.[12] Because most of them present at an advanced stage, they have an extremely poor prognosis.[12] Thus, their early detection is crucial. The tools for this purpose comprise of a combination of imaging and sampling techniques. The imaging methods include ultrasound (US), computed tomography (CT), magnetic resonance imaging (MRI), retrograde cholangio-pancreatography (ERCP), and endoscopic ultrasound.[3]

Among the sampling methods, endoscopic ultrasound-guided fine needle aspiration cytology.

(EUS-FNAC) has established itself as the method of choice because of its ability to target lesions which are much smaller than those that can be targeted by the more traditional imaging techniques like the transabdominal ultrasound or CT.[45] Notwithstanding its popularity, the procedure itself requires specialized expertise and the interpretation of the smears can be challenging.[56] Lesions located in the head of the pancreas are aspirated through the transduodenal route while those located in the body and tail of the pancreas are aspirated through the trans-gastric route.[57] Therefore, duodenal and gastric epithelial cells in addition to gastrointestinal mucin are commonly seen on smears; creating potential pitfalls.[57] Besides, the cytological diagnosis of adenocarcinomas may be tricky, especially in cases of well-differentiated adenocarcinomas.[8] A number of articles highlight these challenges with some suggesting the use of specific sets of diagnostic criteria to improve diagnostic accuracy.[78910]

This is a retrospective study detailing the analyses of FNAC smears and available cell block preparations from 59 cases reported as pancreatic adenocarcinomas using a range of semiquantitative scoring methods wherever applicable.

MATERIALS AND METHODS

A total of 59 cases reported as pancreatic adenocarcinomas on FNAC at the cytopathology laboratory of a tertiary care hospital between December 2016 and September 2019 were analyzed retrospectively.

The FNAC smears were prepared from aspirates obtained from pancreatic lesions using endoscopic ultrasound-guided fine needle aspirations. Transduodenal or transgastric approaches were used to target the lesions depending upon their locations at the head or body/tail of the pancreas, respectively. Papanicolaou stain was performed on smears fixed in 95% alcohol, while May–Grunwald–Giemsa (MGG) staining was performed on air-dried smears. In 32 cases, rapid on-site evaluations (ROSE) was performed by a cytopathologist. In a majority of cases, needle rinses were prepared in containers containing 10 ml of a liquid-based cytology fluid (Thinprep ® Hologic).

For preparing cell blocks, the needle rinses were centrifuged and the pellets added to a tube containing 200 μl of fresh plasma to which a pinch of thrombin was added. The tube was left to stand till a clot developed. This clot was placed in a cassette which was transferred to a container of formalin for preparation of paraffin block from which the cell block sections were prepared.

Light microscopic reviews were performed for pancreatic FNAC smears, lymph node aspirates, and cell block preparations. Clinical records and histopathology reports were reviewed for acquiring follow-up data.

Evaluation of pancreatic FNAC smears: The FNAC smears for evaluated for various cytomorphological features. These have been tabulated in Table 1 along with a description of the methods of evaluation of the individual features.

Table 1

The cytomorphological features that were evaluated and the methods of these evaluations

The evaluated features	Methods of evaluating the features
Nonmalignant cells	Pancreatic ductal cells), acinar cells, duodenal epitheliumGastric epithelium, neutrophils, squamous epithelium
Cellularity	1+ = <10% smear area, 2+ = 10 to 25% smear area, 3+ = 25 to 50% smear area and 4+ = > 50% of the smear area.The highest score was assigned to the case.
Necrosis	0=no necrosis, 1+ = less than 10% of smear area, 2+ = 10 to 25% of smear area, 3+ = 25-50% smear area and 4+ = > 50% of the smear area. The highest score was assigned to the case.
Extracellular mucin	Presence or absence of extracellular mucin was noted.
Stromal fragments	0=no stromal fragments, 1+ = less than five stromal fragments and 2+ = More than 5 stromal fragments.The highest score was assigned to the case.
Cell arrangement	Evaluation of smears for the presence of tissue fragments, sheets, cohesive clusters, dyscohesive clusters, isolated cells.
Nuclear arrangement	Evaluation of nuclei for overlapping and crowding, deranged polarity acinar pattern, pseudo-stratification of nuclei, feathering of nuclei
Anisonucleosis	Based on comparative size of largest nuclei in an aggregate as compared to smallest nuclei1+: The largest nucleus was less than twice the smallest2+: The largest nucleus was 2 to 3 times the smallest3+:. The largest nucleus was 2 to 3 times the smallest4+: The largest nucleus was more than 4 times the smallest
Nuclear membrane irregularity	1+Shallow clefts were present2+Deep clefts were present
Nucleoli	0: No nucleolar prominence, 1+: Nucleoli visible only at 40 x magnification, 2+ Nucleoli visible at 10x magnification
Chromatin abnormality	The nature of chromatin was evaluated and designated as euchromatic, hyperchromatic or irregularly clumped chromatin
Mitosis	The number of mitosis per 100 malignant cells was estimated from the most mitotically active areas by counting 500 cells and dividing the value by five.
Amount of cytoplasm and additional cytoplasm related cellular characters	The amount of cytoplasm was denoted as scanty, moderate and abundant depending upon the nucleo cytoplasmic (n/c) ratioScanty: n/c ratio >90%, Moderate: n/c ratio 50 to 90%,Abundant: n/c ratio <50%Other cytoplasmic features specifically looked for were: a) malignant cells distended with mucin with sharp cytoplasmic edges (referred to as beehive cells) b) signet ring cells.

Evaluation of peripancreatic lymph node aspirates: In cases where peripancreatic lymph node aspirates were submitted, the smears were evaluated for metastatic carcinoma.

Evaluation of cell blocks: The cell blocks were evaluated for their overall cellularity, the arrangement of the cells, and the cytomorphological characters.

d) Follow-up of cases:

i) Clinical follow-up: The follow-up of the patients were obtained from the clinical records department.

ii) Histopathology follow-up: When available, the histopathology reports of the cases were reviewed for correlations between cytopathology and histopathology diagnoses.

Ethics committee approval

This study was performed according to the guidelines of the of the local ethics committee, which conforms to the “World Medical Association Declaration of Helsinki.” Consent was obtained orally in these cases.

RESULTS

The patients comprised of 20 females and 39 males with an age range between 27 and 86 years and a median age of 63. There were 52 lesions in the head of the pancreas and seven in body of the pancreas. In all cases smears were available for review, whereas in 54 cases cell block sections were was also available for review. All the cases had been diagnosed as ductal adenocarcinomas of the pancreas during initial reporting and following review, these diagnoses remained unchanged.

The cytomorphological findings have been summarized in Table 2.

Table 2

Findings from the analyses of individual cytomorphological characters

Cytomorphological characters	Prevalence of characters amongst cases with percentage*
Non- malignant cells	Pancreatic ductal cells: 23 (39.65%)Acinar cells: 3 (5.17%)Duodenal epithelia: 12 (20.68%)Gastric epithelia: 2 (3.44%)Neutrophils: 3 (5.17%)Squamous epithelia: 4 (6.89%)
Cellularity	1+	2+	3+	4+
	13 (22.41)	15 (25.86%)	20 (34.48%)	10 (17.24%)
Necrosis	1+	2+	3+	4+
	17 (29.31%)	4 (6.89%)	8 (13.79%)	1 (1.72%)
Mucin	Present: 5 cases
Stromal fragments	0	1+	2+
	18 (31.03%)	28 (48.27%)	12 (20.68%)
Cell arrangement	Tissue fragments: 45 (77.86%)/Sheets: 54 (93.10%)Cohesive clusters: 55 (94.82%)/Dyscohesive clusters: 49 (84.48%) Isolated cells: 47 (81.03%)All of the above: 33 (57.89%)
Nuclear arrangement	Overlapping and crowding: 58 (100%)Deranged polarity: 58 (100%)Acinar pattern: 15 cases (25.86%)Pseudo-stratification of nuclei: 6 cases (10.34%)Feathering of nuclei: 5 cases (8.62%)
Anisonucleosis	1+	2+	3+	4+
	14 (24.13%)	24 (41.37%)	16 (27.58%)	4 (6.89%)
Sudden pleomorphism: 4 cases (6.89%)
Nuclear membrane irregularity	1+			2+
	58 (100%)			21 (36.20%)
Nucleoli	1+			2+
	7 (12.06%)			19 (32.78%)
Chromatin abnormality	Only euchromatic: 5 (15.51%)Only hyperchromatic: 3 (5.17%)Euchromatic and hyperchromatic: 49 (84.48%)Irregularly clumped chromatin: 18 (31.03%)
Mitosis	0/100 cells	2/100 cells	3/100 cells	4/100 cells
	25 (43.10%)	23 (39.65%)	9 (15.51%)	1 (1.72%)
Amount of cytoplasm and additional cytoplasm related cellular character	Scanty: 42 (72.41%)Moderate: 44 (75.86%)Abundant: 32 (55.17%)Giant cells: 2 (3.44%)Signet ring cells: 3 (5.17%)Beehive cells: 16 (27.58%)

*The proportion of cases with individual characters has been calculated as a percentage of 58 cases with diagnostically representative material on the smears. (One case without any malignant cell on smears and diagnosed on cell block has been excluded)

Nonlesional cells: The duodenal epithelial cells were arranged in sheets and clusters and were identified based on the uniform spacing of the nuclei, admixture of goblet cells, and luminal brush borders [Figure 1a]. Some sheets of duodenal epithelium showed marked nuclear overlapping and a fair degree of disordered polarity. The identification of a distinctive brush border helped in identifying them as benign [Figure 1b]. The gastric foveolar cells were identified based on their distinctive apical mucin caps [Figure 1c]. Pancreatic ductal cells were the most common of all nonlesional epithelial cells. They comprised mostly of mono-layered clusters of oval to round cells with normo-chromatic monotonous nuclei and moderate amounts of cytoplasm [Figure 1d]. Also seen were benign squamous cells (possibly due to procedure related contamination). A few cases showed numerous neutrophils.

Figure 1

(a) Duodenal epithelial cells seen on a transduodenal EUS-FNAC. The clear spaces (arrow) represent goblet cells. (Pap: 20X) (b) Duodenal epithelial cells. Although, extensive nuclear overlap and a fair degree of disarray of nuclear polarity are noted, the distinctive brush border (arrow) helps in recognizing this cell fragment as benign. (Pap: 20X) (c) Gastric foveolar epithelial cells seen on a transgastric EUS-FNAC. The apical mucin caps (arrow) help in the identification of these cells. (Pap: 40X) (d) Pancreatic ductal epithelial cells with evenly spaced monomorphic nuclei. (Pap: 20X)

Cellularity: In one case malignant cells were absent. A definite diagnosis was possible because of diagnostic material on cell block. 3+ was the most common cellularity score, followed by 2+ and 1+. The least common was 4+.

Necrosis: Necrosis was present in 30 (53.44%) of the 57 cases. In a majority, it was scored as 1+

Extracellular mucin: Extracellular mucin was noted in five cases (8.6%).

Stromal fragments: Stromal fragments were present in 40 cases (68.96%). In 28 cases (48.27%), the stromal fragments were estimated to be 1+ and in 12 cases (20.68%) 2+. There was no apparent correlation between the number of stromal fragments and cellularity.

Cellular arrangement: A wide variety of arrangements were noted as shown in Table 1. The arrangements were tissue fragments, sheets, cohesive clusters, dyscohesive clusters, and individually scattered cells. In 33 (57.89%) cases, all these arrangements were present together in the same sample.

Arrangement of nuclei: Overlapping and crowding of nuclei and deranged polarity were noted in all the cases [Figure 2a]. The other noteworthy observations were acinar arrangements [Figure 2b], pseudo-stratification of nuclei at the edges of clusters [Figure 2c], and feathering [Figure 2d].

Figure 2

(a) A sheet of tumor cells. Although the nuclei are minimally pleomorphic, the orientations of their axes with respect to each other are deranged. (Pap: 10X) (b) Tumor cells are arranged in acini. (Pap: 40X) (c) Nuclei of tumor cells show pseudo-stratification (arrow). (Pap 20X) (d) Tumor cells show feathering at the edges (arrow). (Pap: 20X)

Anisonucleosis: The anisonucleosis was scored as 2+ in most cases followed by 3+ and 1+. The least common anisonucleosis score was 4+. Sudden pleomorphism was noted in four cases [Figure 3a].

Figure 3

(a) Sudden pleomorphism among neoplastic cells (arrow) (Pap: 40X) (b) A uniform population of neoplastic cells with minimal pleomorphism, regular nuclear outlines, euchromasia, and inconspicuous nucleoli. (Pap: 20X) (c) The neoplastic cells, show fairly opened up to irregularly clumped chromatin, irregular nuclear outlines, and nucleolar prominence. Multinucleated giant cells with nuclear moulding (arrow) are also seen. (Pap: 40X) (d) The neoplastic cells form a tight cluster. They show high nucleo-cytoplasmic ratios and hyperchromatic nuclei. (Pap: 20X)

Nuclear membrane irregularities: All cases showed cells with at least 1+ irregularity of nuclear membranes, whereas about 36% showed additional cells with 2+ nuclear membrane irregularities [Figure 3b and c].

Nucleoli: Nucleolar prominence was found in 26 cases. 2+ nucleolar prominence scores were more common than a 1+ prominence score.

Appearance of chromatin: Most cases showed cell populations with variable chromatin characters ranging from euchromatic [Figure 3b and c] to hyperchromatic [Figure 3d]

Mitosis count: In about 43% of cases, the mitosis count was zero per 500 cells. In the remaining cases the mitosis ranged from 2 to 4 per 100 cells.

Cytoplasm: Most cases had a mixture of cells with scanty, moderate, and abundant amount of cytoplasm. A few cases showed multinucleated tumor giant cells with nuclear moulding [Figure 3c]. The phrase “Beehive cells” was used to describe cells with mucin filled cytoplasm with centrally located nuclei and sharp cell margins separating the cells from each other [Figure 4a]. Cells distended with mucin, pushing the nuclei to the periphery including signet ring cells were noted [Figure 4b] in a few cases. However, the number of signet ring cells was not sufficient to warrant a diagnosis of signet ring carcinoma in any of those.

Figure 4

(a) The neoplastic cells are filled with mucin, with centrally placed nuclei and sharp cytoplasmic borders, separating adjacent cells (arrow). These were named “Beehive Cells.” (Pap: 40X) (b) Neoplastic cells with voluminous amounts of mucin distending the cytoplasm including signet ring cells (arrow). (Pap: 20X)

Peripancreatic lymph nodes: In four cases, peripancreatic lymph node aspirates were submitted. Two showed metastatic carcinoma.

Cell blocks: Cell block preparations were available in 54 out of the 58 cases. In one case, the diagnostic material was only present on the cell block and not on the smear. In three other cases, a definitive diagnosis of adenocarcinoma was possible after corroborating the cytomorphological findings on the smears with the morphology available on cell block because the smears showed only a few clusters of neoplastic cells.

In 28 (51.85%) cases with available cell blocks, the malignant cells were present as acini and strips, in 15 (27.77%) they were present as acini, strips and clusters, in eight (14.81%) they were present as acini and single cells and in two (3.70%) they were only present as single cells. Stroma fragments with infiltration by malignant cells were noted in 14 (25.92%) cases. [Figure 5a-d]

Figure 5

(a) Small nests of malignant cells and single cells seen to infiltrate through desmoplastic stroma on cell block preparation. (Cell Block: H and E X 20) (b) A small cluster of malignant cells in the cell block preparation showing nuclear disarray and hyperchromasia. (Cell Block: H and E X 20) (c) Malignant cells with mucin distended cytoplasm exhibiting gland formation. (Cell Block: H and E X 20) (d) A sheet of markedly hyperchromatic malignant cells with scanty cytoplasm. (Cell Block: H and E X 20)

Follow-up

Clinical follow-up data were available in 30 cases and unavailable in 29. A total of 6 patients were staged as II, 1 patient was staged as III, and 23 were staged as IV. All the patients who were staged as II and III were alive at the time of acquirement of clinical data. Out of the 22 stage IV patients, 7 were alive, and 22 patients had expired at the time of acquirement of data. Histopathology follow-ups were available in nine cases; all of which were reported as ductal adenocarcinoma.

DISCUSSION

EUS-FNAC has established itself as the method of choice in diagnosing pancreatic lesions.[345] Amongst pancreatic tumors, adenocarcinomas are the most common.[12] The morphological features of these tumors may show a significant degree of variability and have been described and discussed in literature.[578910] A fair proportion of pancreatic adenocarcinomas is well differentiated and their correct identification on cytology may be challenging.[6811] Variable combinations of cytological features have been proposed as minimum requirements toward rendering this diagnosis.[57810] In the present study, detailed cytomorphological analyses of 59 cases reported as adenocarcinomas at our laboratory were performed.

Aspirates from these tumors present a number of critical issues related to the presence of nonlesional cells. It has been reported that FNAs from adenocarcinomas rarely show normal ductal and acinar epithelium,[5] However, in the present study, almost 40% of cases showed pancreatic ductal and or acinar cells. Most importantly they were present on the same smears as the malignant cells.

The correct identification of benign gastrointestinal epithelium which frequently contaminates these aspirates is necessary to avoid pitfalls. The appearance of duodenal epithelial cells reviewed by us mostly conformed to what has been described in literature[57] though, in a certain number of cases we found clusters and sheets of duodenal epithelia with nuclear crowding and overlapping (in contrast to the usual occurrence of evenly spaced nuclei). Since, these sheets and clusters did not show goblet cells, the recognition of a luminal brush border in such situations was crucial in identifying them as benign. To the best of available knowledge, this finding is yet to be described in literature.

Chi et al.,[12] described five pancreatic lesions, diagnosed as adenocarcinomas on cytology which were subsequently found to be other lesions on histopathological evaluation. On histopathology, they comprised of intraductal papillary neoplasms with various degrees of dysplasia and pancreatic intraepithelial neoplasms. Since, the cellularity in all of them were low, the authors recommend a cautious approach in diagnosing adenocarcinomas in hypo-cellular lesions. We scored the cellularity of the aspirates semi-quantitatively. Approximately, 13% of our cases showed a cellularity of 1+ implying that less than 10% of smear area showed malignant cells in such cases. In one case, none of the smears showed any malignant cells. The diagnosis of adenocarcinoma could be rendered based upon the findings on cell block sections. In three other cases, only a few clusters of lesional cells were present. Based merely on smear findings these cases would have been diagnosed as suspicious. The presence of diagnostic material on the corresponding cell blocks allowed a definitive diagnosis of adenocarcinomas to be rendered. Preparing cell block sections should be mandatory in pancreatic FNACs because their use increases the number of cases where a definitive diagnosis can be provided.

The three features which are known to occur in almost all pancreatic adenocarcinomas are loss of honeycomb pattern (due to deranged polarity), nucleomegaly, anisonucleosis, and irregular nuclear contours.[578910] In fact, these were present in most cases of well differentiated adenocarcinomas, studied by Lin et al.[8] In a review of four major studies collated by Bellizi et al.,[7] similar views were expressed.

The other features, less frequently encountered but when present help in buttressing the diagnosis of adenocarcinoma are cellular dyscohesion, prominent nucleoli, para-chromatin clearing (or irregularly clumped chromatin), and background necrosis. In fact, in challenging situations various authors have proposed minimum sets of criteria to ensure a correct diagnosis.[8910]

Disordered nuclear polarity was a universal finding in all our cases. None of the previous publications delve into any detailed analyses of arrangement of malignant cells. We found that most cases show a combination of patterns ranging from three-dimensional tissue fragments, sheets and clusters of cells variable degrees of cohesions, and individually scattered cells. Similarly, nucleomegaly was seen in almost all malignant cells. Most cases showed malignant cells with variable nucleocytoplasmic ratios. Interestingly, almost 74% of cases showed at least some cells with scanty cytoplasm exhibiting very high nucleo-cytoplasmic ratios. Anisonucleosis has been defined as Lin et al.[8] as a nuclear size which is more than four times (the size of the normal pancreatic duct cells apparently). We quantified anisonucleosis from 1+ to 4+ based on the number of the times the size of the largest nucleus in an aggregate of cells is larger than the smallest. Moreover, in a few cases we were able to denote a phenomenon called dramatic anisonucleosis where a nucleus with severe pleomorphism was found among a sheet or cluster of more mundane looking nuclei. Nuclear membrane irregularities were found in all of our cases and the degree of irregularity was variable. Lesional cells with both irregular and regular nuclear outlines were often noted in the same aspirate. Not all pancreatic adenocarcinomas show prominent nucleoli. Nucleoli were visible in 45% of our cases in concurrence with what has been previously described in literature.[89] Brisk mitosis is not a particularly noteworthy feature of these tumors.[789] We quantified mitosis through a count on 500 cells averaged out on 100. In most cases the mitotic count was less than 1 per/100 cells. In only one case, there were more than 4 mitoses per 100 cells. Similar to some of the other nuclear features the malignant cells showed a range of chromatin related characters. Irregularly clumped chromatin, which when present is a strong indicator of malignancy,[8910] was found in about a third of our cases. Most cases showed cells with both euchromatic and hyperchromatic nuclei. According to the review by Belizzi et al., 57% of adenocarcinomas show necrotic material. Necrosis was noted in 52% of our cases. In most, it occupied less than 10% of smear area.

Besides the above, a number of other features were assessed. Background mucin with the tinctorial characters of gastrointestinal mucin was noted in five cases. Presence of atypical cells with abundant mucin may pose an interpretative challenge because mucinous neoplasms and intraductal papillary mucinous neoplasm with variable degrees of atypia may present similar features. A careful review of imaging studies along with fluid biochemistry for carcino-embryonic antigen should be helpful in such situations.[13]

The diagnosis of signet ring cells adenocarcinoma requires a signet ring cell content of more than 50% of the malignant cells on tissue specimens. While three of the cases in the present study showed signet ring cells, they were not present in the required abundance to suggest a diagnosis of signet ring adenocarcinomas. One feature often seen in pancreatic adenocarcinomas is the presence of cells distended with mucin with sharp cytoplasmic outlines. We refer to them as beehive cells. These were noted in about 28% of our cases.

It is a practice in our laboratory to prepare cell blocks from aspirates from all solid pancreatic lesions. Tumors with heavy desmoplastic reactions may yield hypocellular aspirates but on cell blocks they may show tumor cells infiltrating through desmoplastic stroma.

We were able to find follow-up data in more than half our cases. All cases w diagnosed at stage II, underwent a Whipples procedure and histopathological evaluations in all of them showed adenocarcinomas.

The study creates a detailed knowledgebase of cytomorphological analyses of pancreatic adenocarcinomas. Findings of the semi quantitative analyses of a number of features like anisonucleosis, nucleo-cytoplasmic ratios, nuclear irregularities, necrosis, and cellularity can serve as reference points for future studies. In addition, we show how cell block sections might help us in rendering a definitive diagnosis of malignancy when smear findings are equivocal.

Declaration of patient consent

The authors certify that they have obtained all appropriate patient consent forms. In the form, the patient(s) has/have given his/her/their consent for his/her/their images and other clinical information to be reported in the journal. The patients understand that their names and initials will not be published and due efforts will be made to conceal their identity, but anonymity cannot be guaranteed.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.